Add tds.get_field_statistics()

[mgdenno]

We currently have tds.get_unique_field_values("field_name") that returns the unique values which is useful for categorical data. I think it would be useful to have something similar for continuous fields such as reference_time, datetime, etc.

[mgdenno]

The could include number of distinct values which could help with #175 although probably not the solution we are looking for.